Research

Current Research

As you may have guessed from the intro page and the figure above, my main research focus is on the chemical abundances of red giant branch (RGB) and asymptotic giant branch (AGB) stars in Galactic globular clusters. Along with my previous and current collaborators: Caty Pilachowski (advisor), Chris Sneden (undergrad advisor), Bob Kraft, Inese Ivans, Jon Fulbright, and Michael Rich, I intend to help build on the present database of globular cluster data by taking advantage of the Hydra multifiber spectrographs on the 3.5 meter WIYN telescope at KPNO and the Blanco 4 meter telescope at CTIO. My dissertation project involves determining abundances of O, Na, and many other elements in hundreds of stars in the globular cluster ω Centauri, an unusual cluster that has an intrinsic metallicity spread of more than a factor of 10! Additional projects include determining the chemical composition of red giant stars in the Galactic bulge and infrared photometry of Galactic globular clusters.

Why Globular Clusters?

The left image is a photograph of the globular cluster M3 (S. Kafka and K. Honeycutt) and the right image is an annotated color-magnitude diagram of M3 (Kraft 1994).

Globular clusters are an interesting group of objects because they are very old (several billion years old), have a wide variety of stars at different evolutionary states running from the main sequence to the AGB (or beyond), and are anywhere from 10 to more than 100 times more metal poor than the Sun. All of these stars are at approximately the same distance and have other similar properties such as reddening (assuming small differential reddening effects), radial velocity, metallicity, etc. This makes globular cluster stars (and likewise open cluster stars) easier to analyze than field stars, where some of these properties are usually not as accurately determined.

Gather Photons And Then...

In order to obtain information about these stars, we spread the light out into its individual wavelength components using either a prism (see picture at top of page) or (in most cases) a diffraction grating. This creates a spectrum that is captured on a CCD chip and recorded for analysis (see image below).

Each atom and molecule present in the atmospheres of these stars has a specific wavelength signature that corresponds to various electron transitions so when an atom absorbs a photon, the light intensity at a certain wavelength decreases and you see absorption features like those above. By comparing the observed spectra to a known list of reference lines, it is possible to determine which element or molecule is in the star's atmosphere.

Determining Abundances

Now that we know which elements or molecules are present, we can measure the depth and width of the lines to determine how much of each element/molecule is present. To do this, you need: the observed spectra, a model atmosphere, and a linelist that has basic information about each line you intend to measure (typically something like wavelength, excitation potential, oscillator strength, etc.). If the physical parameters of the star are known, namely effective temperature and surface gravity, then you can calculate how many atoms are necessary to give the observed spectral feature. Several groups have developed their own programs to do the necessary calculations to determine abundances and I use the one developed by Chris Sneden called MOOG. One can determine abundances by either generating synthetic spectra (example shown below) and matching it to the observed or measuring equivalent widths.

Synthetic spectral analysis has the advantage of being more "hands on" and enables one to visually fit lines that are blended with nearby features. The equivalent width method is faster, but it can also handle blending and hyperfine/isotopic structure. The solid, dotted, and dashed lines above correspond to different abundances of Al for a metal-poor and metal-rich star in the globular cluster ω Centauri. The solid lines are the best fit abundances while the dotted lines are varied by +/- 0.30 dex (about a factor of 2). For comparison, the dashed lines show Al abundances decreased by a factor of 100,000 (i.e., if no Al were present in these stars). Abundance values are typically given as ratios relative to Fe (or H) and normalized to the abundance in the Sun (notation: [X/Fe] or [X/H]).

Synthesis Of The Elements

Now that we've calculated abundances for the various elements the crucial next step is deciding how it all fits together to give a clear picture of the chemical evolution of the Galaxy. Since the universe consists overwhelmingly of hydrogen and helium, the assumption here is that many of the heavier elements (which you are made of) were synthesized inside earlier generations of stars. How are they created? Most elements can be created in several different ways, but the most common processes are: quiescent core and shell burning, explosive burning, proton capture, and neutron capture.

NeNa and MgAl Proton Capture Cycles

Main sequence stars of about a solar mass and less convert hydrogen to helium predominantly via the PP (proton-proton) chains, while stars a bit more massive than the Sun use carbon, nitrogen, and oxygen as catalysts to produce the same thing through the CNO cycle. After stars exhaust their central hydrogen sources, they become cooler, more luminous, and larger as they once again fuse hydrogen to helium, but this time in a CNO burning shell surrounding an inert helium core. This is the beginning of the red giant phase. Post main sequence evolution differs depending on initial mass, but the general idea is that stars will continue to fuse heavier and heavier elements in both the core and outer shells until they reach Fe, the most tightly bound nucleus. A star's terminating composition depends strongly on its mass, so smaller stars like the Sun fuse up to carbon and then become white dwarfs, while more massive stars can potentially make it to Fe, but blow themselves up as type II supernovae.

Well that explains some of the elements found in the universe, but what about everything else? Elements heavier than Fe are produced by successive neutron captures on seed nuclei (usually Fe or other Fe-peak nuclei) in either the slow (s) process or rapid (r) process. The s-process occurs when neutrons are captured by the seed nuclei on a timescale long compared to the ß-decay rate of the resulting nucleus and the r-process occurs when neutrons are captured on a timescale short compared with ß-decay rates. Where do these neutrons come from? The two most likely sources are the lower temperature ¹³C+⁴He→¹⁶O+n and higher temperature ²²Ne+⁴He→²⁵Mg+n reactions. The s-process operates in low mass (1-3 solar mass) AGB stars by exposing seed nuclei to neutrons created (in radiative burning conditions) in a ¹³C pocket in between thermal pulses. This process creates many of the elements up to lead and bizmuth, where the path is halted due to rapid ß-decays. In contrast, the r-process operates in more extreme situations associated with the deaths of high mass (~8-10 solar mass) stars, where temperatures and neutron densities are very high. This enables the neutrons to be captured more quickly than the nuclei can ß-decay, resulting in element synthesis all the way up to uranium!

What are the possible polluting mechanisms? As you can guess from the two sources mentioned above, the most likely culprits are supernovae (massive stars that blow up) and AGB stars (which eject material via stellar winds as they age). Are these reasonable assumptions? It would seem so. For very metal poor (i.e. old) field stars ([Fe/H]<-2.5), there is an overabundance of r-process relative to s-process elements, suggesting that whatever is creating the r-process elements is doing so on short time scales. Something must have polluted these lower mass stars we are observing. This leaves the main source as high mass stars, which live only millions instead of billions of years. Likewise, the short-lived, radioactive element technetium has been detected in AGB stars. Since the half life of the technetium isotope is only about 100,000 years, clearly it is being created in the observed star and not an earlier generation.

Orderly Disorder

So what do we see in field and globular cluster stars and where does this lead? The elements are not created at random and there are distinguishable correlations and anticorrelations among many elements, especially the "light" metals (i.e. carbon-aluminum). Nearly all red giant stars exhibit the signs of internal mixing as their CNO burning shells and convective envelopes cause C to decline, N to enhance, and the ¹²C/¹³C ratio to approach the equilibrium value of 4. The C+N+O sum is also more or less constant at all RGB luminosities, which is expected if it were indeed the CNO cycle causing these abundance effects. What about oxygen and the other light elments that are sensitive to proton capture nucleosynthesis in these low mass, metal-poor stars? Here is where the story really gets interesting. In many globular cluster RGB stars, there is a clear anticorrelation of O versus Na, but this is not seen in field stars and the degree of oxygen depeltion and Na enhancement varies even among clusters of similar metallicity.

O versus Na anticorrelation for several globular clusters (Ivans et al. 2001).

The most prominent example is in comparison of the two clusters M3 and M13. Despite having nearly identical metallicity ([Fe/H]=-1.50), the most luminous stars in M13 show severe oxygen depeltion ([O/Fe]<-0.40) coupled with a large percentage of high Na, high Al stars. A curious feature about these two clusters is that they have strikingly different horizontal branch (HB) morphologies, the so called "second parameter" (the first parameter is metallicity). M3 has a nearly uniform distribution of red horizontal branch (RHB), RR Lyrae, and blue horizontal branch (BHB) stars, while M13 has almost exclusively BHB stars. Is this difference the same one that drives the abundance anamolies in these two clusters? That is unknown at the moment, but it has been suggested that the different HB's may be due to either age or different ab initio He abundances (with M13 being either older or higher in He content). Recent data indicate that cluster mass could also play a significant role because more massive clusters have bluer horizontal branches and show more stars with low [O/Fe] ratios and high [Na/Fe] and [Al/Fe] ratios.

So were these abundances "built in" to the gas from which these stars formed or have they simply evolved this way? The relationship among the various light elements strongly hints that they were created (or destroyed) during proton capture cycles operating at temperatures in excess of 40x10⁶ K. A big question remaining is where these events transpired. The two main processes proposed are in situ mixing occurring in the current generation of RGB stars and hot hottom burning that occurred at the bottom of the convective envelope in an earlier generation of intermediate mass AGB stars. Unfortunately, neither source is able to fully explain the present observations. This story may in fact hinge on other factors as well, such as nuclear cross sections, reaction rates, and detailed stellar evolutionary physics, which are equally as important as observations.